Re: New PowerPC 7400 Code

[BRENDAN CHRISTOPHER YOUNGER]

Unfortunately, it doesn't.  I've also since discovered that it doesn't
work for overlapping source and destination.  (An oversight on my
part.)  Fortunately, it shouldn't be difficult to rectify either of
these failings and I expect to have revised C code done as soon as I
have time.  (Probably this weekend.)  Assembly code might take a little
longer, but before I redo parts of that, I'd be interested to know what
the relative calling frequencies are for this function.  (i.e., how long
are the limb arrays?)
And finally, the code I sent assumes 32-bit limbs and no nail bits.  I
doubt there are plans to change either of these details in the future,
but if there is, I might be able to hack out a general function for
either case.

Brendan Younger

----- Original Message -----
From: Kevin Ryde <address@hidden>
Date: Monday, October 7, 2002 10:21 pm
Subject: Re: New PowerPC 7400 Code

> Brendan Younger <address@hidden> writes:
> >
> > It takes advantage of the AltiVec instructions to give a
> > speedup of 2x for non-bandwidth bound operations and a speedup of
> > about 30-40% for bandwidth bound operations.
> 
> Do you allow for operands not aligned on 16-byte boundaries?  mpn
> operands will have only 4-byte alignment in general.
>